2018 年 59 巻 3 号 p. 380-385
The mechanical stability of retained austenite in a Cr-Ni weld metal in a welded joint under a service load is related to the initial amount of retained austenite and the type of welded joint. The welded joint used is the over-matching type, i.e., the tensile strength of the weld metal is higher than that of the base metal. For this welded joint, when the initial amount is less than 7.5%, the retained austenite is stable enough to remain in the weld metal over the whole tension process. However, when the initial amount is 12–27%, the retained austenite is metastable, and a large amount of it has transformed into martensite at the applied stress level of 0.2% proof stress. At that stress level, 12% of the retained austenite is still present. As the applied strain further increases, the retained austenite continues to transform, but 7.5% of it is still available after the failure of the welded joint.
When metastable austenite is imposed by applied stress or strain, it will transform into martensite, accompanied by dilatation and strengthening. This phase transformation is denoted as stress- or strain-induced martensite transformation. This type of phase transformation is generally used to improve mechanical properties, such as strength and uniform elongation, in conventional TRIP steel.1–5) A new Cr-Ni weld metal of 980 MPa grade with a microstructure of martensite and retained austenite (RA) was developed based on the concept that martensite guarantees high strength and retained austenite elevates mechanical properties and toughness as well as preventing hydrogen embrittlement. It has been confirmed that the retained austenite in the Cr-Ni weld metal enhances the mechanical properties6–8) and fracture toughness9,10) through stress- or strain-induced martensite transformation and that it effectively inhibits hydrogen embrittlement by trapping diffusive hydrogen.11)
In practical service, to ensure the above benefits of retained austenite, there must be an appropriate amount in the weld metal. As we know, besides applied stress and strain, temperature drop is another factor which can make retained austenite transform into martensite. The thermal stability of retained austenite in the Cr-Ni weld metal has been investigated elsewhere.12) It was found that if the initial content of retained austenite is less than 20%, even if the service temperature reduces from 25℃ to −196℃, its amount remains constant. This demonstrates that within the temperature range of 25℃ to −196℃, temperature variation does not affect the amount of retained austenite when its initial amount is less than 20%. However, one question remains: without a service load, Ref. 12) indicates that a maximum of 20% retained austenite is available; but under a service load, retained austenite probably partially or completely transforms into martensite by applied stress or strain, and then how much austenite is retained in practical service is unclear. Actually, this concerns the correlation of the amount of retained austenite with the applied force, i.e., the mechanical stability of retained austenite. In this study, our focus is on the mechanical stability of retained austenite in a multi-pass Cr-Ni weld metal in a welded joint under a service load. By investigating the variation in the amount of retained austenite against applied stress and strain, the maximum available amount of retained austenite under a service load is revealed. In practical service, once a failure occurred in a welded joint, fracture is expected to be in the base metal instead of weld metal. Based on this conception, the weld metal is generally designed to have larger strength than that of base metal (i.e., over-matching welded joint). Over-matching welded joint is the most common type, and it is also applied in this study.
HT980 steel plate was used as the base metal. Cr-Ni solid welding wire was developed for this study. The chemical compositions of the base metal and welding wire are given in Table 1. Arc welding was performed by five passes on the HT780 steel plate (25 mm thickness) with an X-type groove under pure Ar shielding gas. The heat input is 19–20 kJ/cm (current: 380–400 A; voltage: 27–30 V; welding speed: 35 cm/min).
| material | C | Si | Mn | P | S | Cr | Ni | Mo | Nb | B |
|---|---|---|---|---|---|---|---|---|---|---|
| base metal | 0.13 | 0.32 | 1.43 | 0.01 | 0.001 | 0.02 | 0.03 | 0.55 | 0.03 | 0.0012 |
| welding wire | 0.049 | 0.61 | 0.71 | 0.016 | 17.7 | 7.13 |
Three types of tensile specimen (tensile specimens of the base metal, weld metal, and welded joint) were machined from the welded joint as shown in Fig. 1. A round tensile specimen was used for the base metal and weld metal. The axis of the tensile specimen of the base metal is vertical to the weld bead. Its diameter is 4 mm, and the gage length is 20 mm. The weld metal tensile specimen was machined from the up and down weld metal along the weld bead as shown in Fig. 1(b). Its diameter is 6 mm, and the gage length is 30 mm. Plate tensile specimens of the welded joint with a width of 20 mm, thickness of 3 mm, and gage length of 30 mm were machined as shown in Fig. 1(a). The weld metal is at the center of the plate specimen. The specimens of the base metal and weld metal were used for achieving their mechanical properties, and the specimens of the welded joint were used for investigating the mechanical stability of retained austenite in the multi-pass weld metal.
Preparation of welded joint and tensile specimens: (a) plate tensile specimen of welded joint and round tensile specimen of base metal and (b) round tensile specimen of weld metal (machined from up and down weld metals along the weld bead). D, diameter; GL, gage length.
Nominal stress versus nominal strain curves of base metal, weld metal, and welded joint.13)
Local strains in the base metal (BM), heat-affected zone (HAZ) and weld metal (WM). Circles 1 and 2, located in the BM; circles 3 and 4, in the WM; circle 5, in the HAZ.13)
Loading cycles below 0.2% proof stress. Loading from 0 MPa to a certain stress level and then unloading to 0 MPa; total seven cycles.
All tensile tests were performed at room temperature and at a crosshead speed of 0.01 mm/sec. The variation in the amount of retained austenite in the multi-pass weld metal against applied force was measured within two loading ranges with two plate specimens of the welded joint―one specimen for the elastic range (from the initial state to the 0.2% proof stress) and the other for the plastic range (from the initial state to the failure of specimen). The plate specimen was tensioned from the initial state (0 MPa) to a certain stress (or strain) level, followed by unloading to 0 MPa, and then the amount of retained austenite was measured. This process was repeated several times until attaining the 0.2% proof stress for the former specimen or until the failure of the specimen for the latter specimen. The amount of retained austenite was measured by X-ray diffraction (XRD). The XRD measurement was described in detail elsewhere.12)
The weld metal is composed of five passes. Because the dilution ratio of the base metal is different for each pass (the first pass, i.e., the center of the weld metal, has the largest dilution ratio), the chemical composition is not uniform. The chemical composition of five places was measured. The five locations are on the y axis (cf. Fig. 1(a)). The center of the weld metal (half the thickness of the weld metal) is taken as the origin of the y axis. The coordinates of the five locations and the corresponding chemical composition are given in Table 2. Table 2 demonstrates that the chemical composition along the thickness direction, especially Ni and Cr, is not uniform. The Ni and Cr contents at the center (y = 0) are the lowest.
| y (mm) | C | Si | Mn | Ni | Cr | Mo |
|---|---|---|---|---|---|---|
| 10 | 0.054 | 0.74 | 0.83 | 6.43 | 14.33 | 0.036 |
| 5 | 0.066 | 0.67 | 0.95 | 5.27 | 11.67 | 0.13 |
| 0 | 0.080 | 0.61 | 1.03 | 4.46 | 9.87 | 0.20 |
| −5 | 0.065 | 0.70 | 0.89 | 5.70 | 12.69 | 0.094 |
| −10 | 0.055 | 0.72 | 0.85 | 6.25 | 13.85 | 0.052 |
Two plate tensile specimens (specimen 1 and specimen 2 shown in Table 3) were used to evaluate the mechanical stability of retained austenite within the elastic region and plastic region, respectively. The initial amounts of retained austenite of five points on the y-axis (cf. Fig. 5 and Fig. 7) were measured. The results are given in Table 3. The locations near the center of the weld metal (specimen 1: y = 2.0 mm and y = −1.3 mm; specimen 2: y = 1.3 mm and y = −1.7 mm) have low amount of retained austenite. It has been shown that the amount of retained austenite is related to the chemical composition, especially Ni and Cr contents: low Ni and Cr contents corresponding low amount of retained austenite.6) This is because Cr and Ni are the elements elevating the stability of austenite. Elevated thermal stability by increasing the amount of Cr and Ni inhibits the phase transformation from austenite to martensite. Low Ni and Cr contents cause the center of the weld metal to have the lowest amount of retained austenite.
| y (mm) | initial RA in Specimen 1 (%) |
y (mm) | initial RA in specimen 2 (%) |
|
|---|---|---|---|---|
| 6.7 | 13.9 | 7.6 | 18.9 | |
| 2.0 | 5.5 | 3.8 | 16.9 | |
| −1.3 | 4.9 | 1.3 | 7.0 | |
| −4.2 | 27.4 | −1.7 | 6.8 | |
| −7.4 | 17.0 | −7.3 | 21.6 |
RA, retained austenite
Specimen 1, for elastic region measurement; specimen 2, for plastic region measurement
Variation in the amount of retained austenite corresponding to the loading histories shown in Fig. 4.
Loading cycles until failure of the tensile specimen. Loading from 0 MPa to a certain stress level and then unloading to 0 MPa; total four cycles.
Variation in the amount of retained austenite corresponding to the loading histories shown in Fig. 6.
The mechanical stability of retained austenite in a welded joint is dependent not only on the retained austenite itself, but also on the deformation performance of the welded joint. The deformation of the welded joint used in this study has been investigated in detail elsewhere by tension test.13) The obtained results in Ref. 13) directly correlate with the variation in the amount of retained austenite in the weld metal. The results are briefly stated as follows.
The tensile properties of the base metal and weld metal were evaluated, and the nominal stress versus nominal strain curves (S-S curves) are shown in Fig. 2. The 0.2% proof stress ($ \sigma_{0.2}$), tensile strength (TS), ratio of $ \sigma_{0.2}$ to TS ($ R = \sigma_{0.2}/{\rm TS}$), and uniform elongation are summarized in Table 4. There is great difference between the base metal and the weld metal in the shape of the S-S curve. The weld metal has a much larger gradient than the base metal within the range from the yielding point (denoting as ${\sigma _{0.2}}$ in this study) to the maximum stress point. It also has smaller R values than the base metal. The gradient and R value are parameters reflecting the work-hardening ability―the higher the gradient (or the smaller the R value), the stronger the work-hardening ability. The larger gradient and smaller R values indicate that the weld metal has a much stronger work-hardening ability than the base metal. The increased work-hardening ability is attributed to the stress- or strain-induced martensite transformation from the retained austenite.7)
| material | σ0.2 (MPa) | TS (MPa) | R = σ0.2/TS | uniform elongation (%) |
|---|---|---|---|---|
| base metal | 1035 | 1059 | 0.977 | 5.4 |
| weld metal (up) | 419 | 1202 | 0.349 | 5.5 |
| weld metal (down) | 461 | 1216 | 0.379 | 5.5 |
| welded joint | 889 | 1043 | 0.852 | 5.4 |
σ0.2, 0.2% proof stress; TS, tensile strength; R, ratio of σ0.2 to TS
Table 4 shows that the weld metal and base metal have different yield and tensile strengths. If classifying the welded joint used according to the tensile strength level, it is an over-matching welded joint (i.e., TS of the base metal < TS of the weld metal).
The welded joint is composed of the base metal (BM), heat-affected zone (HAZ) and weld metal (WM). Previous research13) shows that the HAZ has the lowest strength. The difference in strength levels in the three parts determines that the deformation performance of the whole welded joint is dependent on the combination of the three parts instead of single part. Tension test on the welded joint was performed, and the stress versus strain curve is also shown in Fig. 2. The strain is an average value over the gage length of 50 mm (including the BM, HAZ and WM) measured by an extensometer.
Necking occurred in the base metal. As necking proceeded, the plate tensile specimen fractured in that location. It is noted that the extensometer used was taken off before the final failure in order to prevent it from the shock of the fracture, and thus, the S-S curve of the welded joint in Fig. 2 is only part of the whole tension process. The $ \sigma_{0.2}$, TS, R and uniform elongation of the welded joint are also summarized in Table 4. The $ \sigma_{0.2}$ and R of the welded joint are larger than those of the weld metal, and smaller than those of the base metal. Apparently, this is the result of the compromise of the BM, HAZ and WM. Necking and final failure in the base metal make the base metal control the tensile strength of the welded joint, which has been verified from the fact that the TS of the welded joint is almost the same as that of the BM. The difference in the fracture elongation between the welded joint and the BM shouldn't be great. However, Fig. 2 shows that the ductility of the welded joint seems to be much inferior to that of BM. This misleading is caused by the two facts: ① the S-S curve of the welded joint does not cover the whole tension process; ② the S-S curve of the welded joint was obtained from the plate specimen while that of the BM was from the round specimen.
The nominal strain used for the S-S curve of the welded joint in Fig. 2 reflects the combination of the BM, HAZ and WM. It does not reveal the strain evolution in each part. Local strains in the three parts were investigated by two-dimensional digital image correlation (2D-DIC) in Ref. 13), and the results are given in Fig. 3. There are five circles: circles 1 and 2 are, respectively, in the left and right sides of the BM; circles 3 and 4 are in the center and down parts of the WM, respectively; circle 5 is in the HAZ. Average strains measured with 2D-DIC within the five circles are compared with the average strain over the gage length of 50 mm evaluated by an extensometer. The ordinate and abscissa represent the local strain in a specific place and the whole strain across the welded joint (BM, HAZ and WM), respectively. The data above the diagonal line (dotted line) mean that the local strains are larger than the average strain of the welded joint. In other words, strain concentrations occurred in these local sites. Figure 3 indicates that only the BM in the right side (i.e. circle 2) has no strain concentration during the whole tension process. Strain concentration in the HAZ (circle 5) is the most serious. At low strain levels, the strain concentration occurred in the center of the WM (circle 3); above ～0.02, the strain in this region is the lowest in the welded joint.
Figure 3 also shows a tendency that when the engineering strain of the welded joint (abscissa) exceeds 0.07, the local strains of circles 1–5 remain constant. As mentioned before, necking occurred in the base metal. After necking, deformation began to concentrate in that position, and other parts stopped continuing deformation. This causes the local strains in Fig. 3 keep constant above the strain of 0.07. Because the necking position is beyond the area of 2D-DIC measurement, strain concentration in that region is not reflected in Fig. 3.
3.3 Mechanical stability of retained austenite 3.3.1 Stress-induced martensite transformation below yield strengthThe mechanical stability of retained austenite in the weld metal of a welded joint can be assessed using two types of tensile specimen―a pure weld metal specimen or a welded joint specimen (cf. Fig. 1). The former type is machined only from the weld metal. The deformation in it is determined only by the weld metal itself. The effect of mismatching the BM, HAZ and WM is not involved. For this reason, the variation in the amount of retained austenite in the tension process reflects the mechanical stability of retained austenite only in the weld metal, not in a welded joint. In Refs. 7) and 14), this type of specimen was used to evaluate the mechanical stability of retained austenite in the Cr-Ni weld metal. In this study, our focus is on the performance of retained austenite in a welded joint under a service load, and thus a plate tensile specimen of the welded joint shown in Fig. 1(a) was used.
Suppose the service load is not fixed: it varies within the range of 0 to 0.2% proof stress of the welded joint ($ \sigma_{0.2}$). Figure 4 shows the loading history.
The plate specimen was tensioned to a certain stress level from the initial state and then unloaded. In the tension process, the retained austenite partially transformed into martensite. Because the plastic deformation was extremely minor, this phase transformation is regarded as the stress-induced martensite transformation. The amount of retained austenite after this tension process was measured by XRD. This process was repeated seven times. The whole experimental process is briefly as follows (cf. Fig. 4): (step 1) the specimen was tensioned to the stress level of $0.53 \sigma_{0.2}$ from the initial state, followed by unloading and XRD measurement (i.e., 0 MPa → $0.53 \sigma_{0.2}$ → 0 MPa → XRD measurement); (step 2) 0 MPa → $0.53 \sigma_{0.2}$ → 0 MPa → XRD measurement; (step 3) 0 MPa → $0.53 \sigma_{0.2}$ → 0 MPa → XRD measurement; (step 4) 0 MPa → $0.65 \sigma_{0.2}$ → 0 MPa → XRD measurement; (step 5) 0 MPa → $0.78 \sigma_{0.2}$ → 0 MPa → XRD measurement; (step 6) 0 MPa → $0.92 \sigma_{0.2}$ → 0 MPa → XRD measurement; (step 7) 0 MPa → ${\sigma _{0.2}}$ → 0 MPa → XRD measurement.
The initial amount of retained austenite on the five locations (points A–E in Fig. 5) is given in Table 3 (Specimen 1). The amount of retained austenite against the maximum stress level of each step is given in Fig. 5. It should be noted that the amount of retained austenite is the average value over a horizontal line passing the corresponding point (A–E) with a length of 1 mm.
The initial amount of retained austenite (applied stress = 0) along the y axis is not uniform. The center of the weld metal (B and C) has the lowest value (～5%). A, D, and E points have a higher initial amount of retained austenite. As the stress level increases, the retained austenite with a higher volume fraction (A, D, and E) continuously transforms into martensite. However, when volume fraction is low, for example, B and C, there is almost no change in the amount of retained austenite from the initial state to $ \sigma_{0.2}$, indicating that almost no stress-induced martensite transformation occurs. Figure 5 shows that in practical service, when applied stress attains $ \sigma_{0.2}$, a large amount of retained austenite has transformed into martensite, and less than ～12% austenite is retained.
3.3.2 Strain-induced martensite transformation until failure of tensile specimenAfter the yield point, plastic deformation becomes dominant, and the untransformed austenite probably continues to transform into martensite by the plastic strain. After failure of the tensile specimen, how much austenite is still retained is our concern.
A method similar to that in the above section was applied. The specimen was tensioned four times, and the four steps are illustrated in Fig. 6. In step 4, when stress attained the maximum value, necking began to occur in the base metal (the strain is 0.037 at this point).
The amount of retained austenite is plotted against the maximum strain in each tension step in Fig. 7. It is noted that the strain is also the whole strain across the BM, HAZ and WM (GL = 50 mm). The five points (F-J) can be classified into two groups according to the initial amount of RA: low initial amount point (H and I, near the center of weld metal) and high initial amount point (F, G, and J, apart from the center of weld metal). For the former, almost no change in the amount of RA during the whole tension process; for the latter, significant variation in the amount of RA only occurred within the strain range of 0 to 0.01, beyond that range almost no change in the amount of RA. The results indicate that the RA with low initial amount is stable and it will not transform into martensite. The same tendency is also shown in Fig. 5. The RA with high initial amount is metastable.
Chemical composition and grain size are known to be correlated with the stability of austenite. Chemical composition in the weld metal is not uniform. Table 2 shows that the contents of Cr and Ni increase as the distance from the center of the weld metal increases. High Cr and high Ni corresponds to the locations with higher amount of RA. High Cr and high Ni should elevate the stability of RA, but the RA in the locations is metastable. Significant difference in grain size for the RA along the thickness of weld metal is not found according to the microstructure observation. Therefore the Cr/Ni contents and the grain size of RA are not responsible for the low mechanical stability of RA with higher initial amount.
Weld metal is composed of martensite and RA. Soft RA is surrounded with hard martensite. Under the applied force, strain or stress prefers to concentrate in the RA than martensite, inducing nonuniform local strain or local stress. The evolution of local strain in the multi-pass weld metal is given in Fig. 3. It can be seen that the local strains along the thickness of WM are not uniform; the local strain in the weld metal apart from the center is almost three times of that in the center of WM (cf. circle 3 and circle 4 in Fig. 3). In the strain range of 0 to 0.01 (whole strain across the welded joint), the local strain increment is great. When strain exceeds 0.01, there is no local strain change in the center of WM, and the increment becomes small in the WM apart from the center. Within the strain range of 0 to 0.01, the significant strain increment leads to the great decrement in the amount of RA for points F, G and J.
In summary of Figs. 5 and 7, when applied stress attains $ \sigma_{0.2}$, a large amount of retained austenite has transformed into martensite, and at most, ～12% retained austenite is available; after further tensioning, plastic deformation becomes dominant, and strain-induced martensite transformation begins to occur, but the amount of formed martensite is low. After complete failure of the specimen, at most, ～7.5% austenite is retained.
In the present study, the weld joint is an over-matching type. This means that after necking, plastic deformation concentrates in the base metal, and the weld metal ceases to further plastically deform. If the welded joint is an under-matching type (TS of the weld metal < TS of the base metal), because the weld metal has low strength, deformation will concentrate in the weld metal and finally fracture in it. In this case, the retained austenite will experience extremely severe local plastic strain, and it probably completely transforms into martensite. Previous research7) has confirmed that when severe plastic strain imposes on the retained austenite in the Cr-Ni weld metal, it will completely transform into martensite. Obviously, over-matching type is beneficial to maintain the RA.
In the present study, the Cr-Ni multi-pass weld metal has lower yield strength and higher tensile strength than the base metal, i.e., over-matching welded joint. The mechanical stability of retained austenite in such weld metal is summarized as follows:
This study was carried out as a part of research activities for “Fundamental Studies on Technologies for Steel Materials with Enhanced Strength and Functions” by the Consortium of the Japan Research and Development Center of Metals (JRCM). Financial support from the New Energy and Industrial Technology Development Organization (NEDO) is gratefully acknowledged.
